Russell Johnson


  • I would think that you would absolutely want to use the parts that the gage will be measuring. If I’m using a pair of calipers to measue the OD of a key way (with spec limits of .00125″ – .00127″, I don’t care if the calipers will detect a body length that has specs of 1.25″ +/- .05″.

    To get a good part to part variation needed, simply get a…[Read more]

  • When doing a GR&R, you are not wanting to see if your process is making good parts. You want to make sure that your selected gage is capable of distinguishing between a good part and bad part. You want to show that your gage is capable of detecting a bad part. If you have no part variation during the GR&R you will not be able to verify that…[Read more]

  • Matt; I had very limited knowledge of statistics when I began my LSS journey. My degree is in criminal justice. I had basic math in college, and a business math course. My black belt training was a challenge. But, I fell in love with six sigma and statistics and was able to gain a lot of knowledge by reading, reading and reading on the…[Read more]

  • It is always a good idea to look at your data using graph’s. I’ve never had an instance where the AD showed my data to be normal, but my graph showed non-normal. BUT, I have, in the past, been able to look at the data graphically, after determining the data to be non-normal, and find outliers, which caused the distribution to be non-normal. I…[Read more]

  • When Cp and Pp are close to the same value, it is a sign that your process is predictable. When Cp and Pp values differ significantly, you can almost be sure that the process isn’t behaving in a predictable manner. In order for you to be able to predict capability (hence provide a Cpk to your customer), one of the requirements is a stable,…[Read more]

  • I agree with MBBinWI. The body of work is very important when considering a BB candidate. I would want to make sure the candidate has a type of training that includes hands on and classroom type lectures. But, I would also want to see the candidate’s body of work at his organization. What type of projects did he lead? What, if any, break…[Read more]

  • What does your collection plan look like? Are you looking at historical data, or is this data that you are currently collecting? One of the first things we need to do is determine homogeneity of the data. I don’t want to make assumptions because I don’t know much about the data you collected. 2,000 data points could be okay, but, if these 2…[Read more]

  • Accenture, formerly known as The George Group, is excellent. I just completed the MBB training. I would recommend them. Their instructors or top of the line. 8 years, 9 months ago

  • Thanks for the responses! I appreciate the input. I am new to the community and, not having anyone in my plant to bounce this off of, I see where these forums will be of great benefit.

    @kicabcm, I agree that whenever FTY < 100%, we should start looking for improvements. Again, thanks for all the input.

    Russ 8 years, 10 months ago

  • Okay, so say I have 100 units tested and 75 pass. That would be 75% FTY. Now, I take the 25 that failed and rework and retest. Out of the 25 I pass 20. Next, I take the 5 that failed and rework and retest and they all pass. Would I/could I use the Process Yield formula (.75*.80*1)? Or, would I simply say it took 130 tests to pass 100 units,…[Read more]

  • I have a product that is tested on a test stand. What would be the best way to measure first time pass rate, second time pass rate, etc.? For example, I have a product that I test and it passes the first time. […] 8 years, 10 months ago

  • Russell Johnson became a registered member 8 years, 10 months ago